most recent 30 from philosophy.stackexchange.com 2025-08-07T19:22:18Z https://philosophy.stackexchange.com/feeds/question/129311 https://creativecommons.org/licenses/by-sa/4.0/rdf https://philosophy.stackexchange.com/q/129311 10 Max https://philosophy.stackexchange.com/users/73943 2025-08-07T05:46:04Z 2025-08-07T15:37:44Z

<p>We humans observe the natural world and science is about seeing things happen and recording and investigating it. In order to discern how it works, we need to observe it multiple times and compare data to try and understand it.</p> <p>Yet, are there processes in the universe that scientists are pretty sure have only happened once? I’m not taking about the universe itself because we aren’t sure it only happened once).</p> <p><a href="https://plato.stanford.edu/entries/scientific-reproducibility/" rel="noreferrer"><em>Repeatability</em> and <em>reproducibility</em></a> are cornerstones of science. Yet conditions for repeating an experiment or making an observation a second time are never truly identical, because the world is in constant flux and we can never create or encounter identical conditions. It is as Heraclitus said, one cannot step in the same river twice.</p> <p>If something only happened once, can it be called science?</p>

https://philosophy.stackexchange.com/questions/129311/-/129312#129312 6 Jo Wehler https://philosophy.stackexchange.com/users/9174 2025-08-07T06:21:44Z 2025-08-07T06:21:44Z

<ol> <li><p>We observe <em>single</em> phenomena. But science always designs a <em>general</em> theory. Science is not about the single event, it is about finding a general law behind all similar events.</p> </li> <li><p>Of course there are important theories, where until today we only know a single event which falls into the range of the theory. E.g., the theory of biological evolution. And also the theories about the origin of life.</p> <p>But therefore these theories are not non-science, not at all.</p> </li> <li><p>Nevertheless a scientific law must always cover a type of events, not a single specific one.</p> </li> </ol>

https://philosophy.stackexchange.com/questions/129311/-/129313#129313 7 Richard Kirk https://philosophy.stackexchange.com/users/68924 2025-08-07T08:30:04Z 2025-08-07T08:30:04Z

<p>The Big Bang is a single event in our universe. Science is investigating that.</p> <p>We can hypothesise other universes. I am not sure it is helpful to do this unless we can somehow detect them - in which case they overlap with our universe and are not really separate. But people do hypothesise other universes, and it may get them somewhere useful some day.</p>

https://philosophy.stackexchange.com/questions/129311/-/129325#129325 4 Lowri https://philosophy.stackexchange.com/users/68445 2025-08-07T18:36:05Z 2025-08-07T18:36:05Z

<h2>Whether some event only happens once depends on the level of abstraction in the description</h2> <p>Depending on the level of description, all events may only happen once.</p> <p>For example, the event that &quot;X Planck units of time after the Big Bang, neutron N decayed&quot; only happens once.</p> <p>And &quot;the New York Yankees lost the 2024 World Series&quot; only happens once.</p> <h2>Events simpliciter are neither science nor non-science</h2> <p>Events are not categorized as science or non-science. People do not generally speak of an <em>event</em> being science.</p> <p>We instead categorize the approach to inquiry as science or non-science. Or a theory as scientific of non-scientific.</p>

https://philosophy.stackexchange.com/questions/129311/-/129326#129326 21 Ted Wrigley https://philosophy.stackexchange.com/users/40310 2025-08-07T20:56:11Z 2025-08-07T20:56:11Z

<p>Every event only happens once. The purpose of the sciences is to categorize <em>types</em> of events that fall into the same basic models. In other words, If I drop an apple that is a singular event: it happens in one place at one time with one apple. But if I make a model of how <em>any given</em> apple will fall if dropped, then I'm doing science.</p> <p>Basically a scientific model says that we can treat a wide assortment of unique, individual events as though they are the same by removing all of the unimportant, idiosyncratic features that make them unique. For instance, a model of a falling apple will ignore the variety, color, taste, and specific size and shape of the apple, as well as the particular time or location of the event, since none of these attributes (in simplified conditions) are essential to model the event of falling. All of the potentially infinite unique events of dropping apples are hedged or equivocated as the same <em>type</em> of event, in the understanding that this <em>type</em> of event can be replicated in a systematically reliable way (with minute variabilities that we can gloss over).</p> <p>Replication is not the same as repetition. Replication is making a copy, caring only that the copy captures the thing being copied in broad strokes.</p>

https://philosophy.stackexchange.com/questions/129311/-/129354#129354 2 NotThatGuy https://philosophy.stackexchange.com/users/6350 2025-08-07T02:52:56Z 2025-08-07T02:52:56Z

<p>Some particular arrangement and flow of atomic particles may functionally be a one-time event. But science typically doesn't model exact arrangements of particles* as much as it models patterns by grouping objects that behave similarly and similar events.</p> <p>* Unless you're a particle physicist or something. But even there we model similarly-behaving particles.</p> <p>Every apple may be different on the atomic scale. But we can model the category of apples as balls of particles that roughly fit some profile. And so we can draw reliable conclusions about what will happen if you eat or throw such a ball. Some apples are rotten or have worms, and we can also model that as deviations from the &quot;base&quot; apple, and we can model how this changes behaviour.</p> <p>We can also generalise this to include any solid ball of particles, and we can model how those would behave. For example, on Earth, such a ball generally falls to the ground if surrounded by air.</p> <hr /> <p>The above also allows science to investigate events that may not be that similar to anything else we've seen.</p> <p>The Big Bang, for example, is somewhat similar to other expanding things. But even putting that aside, we can look at each galaxy independently to figure out that it's moving away from us. When we trace this backwards, we can conclude that everything was really close together at the &quot;start&quot;. There's more to it than that, but that's one key piece of evidence.</p> <p>To generalise this: we can model individual parts of an event to draw conclusions about the event as a whole, even without considering how similar the event as a whole is to other events. Although similarity to other things helps to validate the conclusions we draw.</p>

https://philosophy.stackexchange.com/questions/129311/-/129360#129360 2 Richard Kirk https://philosophy.stackexchange.com/users/68924 2025-08-07T07:28:03Z 2025-08-07T07:28:03Z

<p>Here is a separate answer which shows how people deal with a real single event in a scientific manner. Millikan demonstrated that charge was quantized by producing oil droplets with a random charge in a spray, and measuring the rate they dropped under gravity, and then the rate when he applied an electric field. The falling rate gave him the size, and hence the weight of the oil drop, and the second measurement gave him the charge. The charge was always multiples of the electron charge, with one exception: one oil drop appeared to have a 1/3 charge.</p> <p>Millikan was a very thorough man. He repeated his experiment many thousand times, and he recorded the one event that did not fit his model. This measurement was remembered when quarks with 1/3 charge became a thing? Had Millikan seen a free quark? People have automated his experiment and repeated it millions of times. No-one has seen another 1/3 charge.</p> <p>It is unlikely Millikan saw a free quark, but there is no easy explanation of how the error came about. Some people claimed Millikan faked his results. Others who repeated his experiments said he was a very careful and honest researcher. In the end we are just left with the record of many measurements of the electron charge, and just Millikan's one measurement of a 1/3 charge.</p> <p>What happened then is the right thing from a scientific point of view. We have an incomplete explanation. We record that Millikan saw a 1/3 charge. We note that no-one else has seen this, and it is very unlikely that a free quark was just floating around Millikan's lab on that particular day. We cannot explain how the result came about. So, we wait. Maybe someone will reproduce what actually happened. More likely, an explanation will come from a different direction as most (all?) scientific fields overlap with others.</p> <p>What</p>

https://philosophy.stackexchange.com/questions/129311/-/129362#129362 1 Philip Klöcking https://philosophy.stackexchange.com/users/17209 2025-08-07T09:56:12Z 2025-08-07T09:56:12Z

<p>While reproducibility and repeatability are cornerstones of the scientific method, they mainly are because science is all about forming theories that are able to produce a hypothesis that is <strong>a prediction</strong>, and test whether this prediction is accurate.</p> <p>Now, can you form hypotheses and make predictions from singular, not reproducible events like the big bang? Yes, of course you can! Those predictions will be very uncertain as a product of <em>abduction</em>, still this singular event (category of one) can be relevant for and accessible by the scientific process.</p>

https://philosophy.stackexchange.com/questions/129311/-/129371#129371 0 J D https://philosophy.stackexchange.com/users/40730 2025-08-07T14:46:57Z 2025-08-07T14:46:57Z

<p>You ask:</p> <blockquote> <p>Can something only happen once? And if so, does that make it non-science?</p> </blockquote> <p><strong>To pile on to the question, reproducibility is only one criterion we appeal to in the labeling of an activity being scientific.</strong> So, yes, we can be scientific about things that happen only one, even if in experimentation, the <a href="https://en.wikipedia.org/wiki/Randomized_controlled_trial" rel="nofollow noreferrer">randomized, double-blind controlled trial</a> is a gold standards of sorts. All sorts of characteristics of scientific activity include <a href="https://en.wikipedia.org/wiki/Peer_review" rel="nofollow noreferrer">peer-review</a>, <a href="https://en.wikipedia.org/wiki/Operational_definition" rel="nofollow noreferrer">operational definitions</a>, <a href="https://en.wikipedia.org/wiki/Mathematical_proof" rel="nofollow noreferrer">mathematical proof</a>, and <a href="https://en.wikipedia.org/wiki/Measurement" rel="nofollow noreferrer">mensuration</a>. History, where events never repeat without abstraction, can be considered a social science where <a href="https://en.wikipedia.org/wiki/Historicity" rel="nofollow noreferrer">historicity</a> is prized and interdisciplinary sciences help to make the discipline more scientific, for instance. All of this difference in scientific methods within different sciences is part of the philosophical <a href="https://en.wikipedia.org/wiki/Demarcation_problem" rel="nofollow noreferrer">demarcation problem</a> Popper raised, and to which modern scholars if science simply concede there is no &quot;one-true&quot; science, but a variety of disciplines that have their own scientific methods. In fact, the sciences may simply have Wittgensteinian <a href="https://en.wikipedia.org/wiki/Family_resemblance" rel="nofollow noreferrer">family resemblance</a>.</p>

https://philosophy.stackexchange.com/questions/129311/-/129372#129372 0 WoJ https://philosophy.stackexchange.com/users/32810 2025-08-07T14:54:30Z 2025-08-07T14:54:30Z

<p>Physicists like to have repeating events that they can match with their model, derive statistics, and validate or invalidate the model. This is the best case.</p> <p>You can also have the weird observation in the sky, recorded by chance, and unexpected. You are not likely to make a model out of it (though I vaguely remember that quite a lot of statistics were done many years ago of some difficult-to-measure events (neutrinos?) where the scientists had <em>two</em> data points (and not a lot of hope to have more)).</p> <p>This can be useful to challenge your existing model, though. Why did you see that, and what in the model prevents that measurement? If you extend this and that, would this new measurement fit? And similar questions.</p> <p>So yes, a single measurement is science, even if not likely to move core knowledge forward.</p>

https://philosophy.stackexchange.com/questions/129311/-/129376#129376 1 alanf https://philosophy.stackexchange.com/users/5759 2025-08-07T15:37:44Z 2025-08-07T15:37:44Z

<p>A scientific theory will have some consequences that are testable and others that are not. For example, the standard theory of how the sun works makes predictions about its magnetic field, how much energy it emits and we can measure and test these predictions. It also makes predictions about the temperature at the sun's core that we can't currently test and may never be able to test.</p> <p>When a test of a scientific theory is performed things can go wrong, such as equipment faults. So if you observe an event that doesn't match a theory and it only happens once that may be a result of a flaw in that particular experimental run. So scientists want the result of some test to be repeatable in the sense that they can control everything that they think is relevant to the test and do this on multiple occasions.</p> <p>An event that only happens once can't be repeated multiple times, but it may be possible to test some prediction of the theory of that event multiple times, e.g. doing multiple measurements of background radiation from the big bang.</p> <p>See &quot;The Logic of Scientific Discovery&quot; by Karl Popper, Section 8.</p>

百度